Seattle, Washington, United States
My name is Akshith Saravanan and I'm a Computer Science student at the University of Washington, Seattle. I am deeply interested in AI/ML, cybersecurity, software engineering, quantum computing, and full-stack web development. Currently, I'm an undergraduate researcher at UW working on studying the proactive information-seeking behavior of AI agents and how this behavior leads to a new class of attacks. My main goal is to help achieve a secure future as people increasingly rely on LLMs in their everyday lives. I'm open to any opportunities at the intersection of AI and cybersecurity. I have a deep understanding of fundamentals such as C/C++ and a strong grasp of system architecture. I have also competed in and secured top positions at various technical competitions. At the DubsTech Datathon, I developed a machine learning model to predict hospital discharge costs and placed 4th internationally. More recently, I was a finalist in the IMC Prosperity 4 Trading Challenge, where my team advanced to the finals, reached the Top 2% of teams before the final rounds, and ultimately finished in the Top 8% globally after competing against 18,803 teams and 30,703 participants worldwide. This experience strengthened my skills in quantitative analysis, algorithmic trading, and data-driven decision making under dynamic market conditions. Through my various internship experiences, I have learned how to write clean code, deploy products, lead teams, and collaborate at scale. I have demonstrated strong skills in AI/ML, Java, web development, and leadership. I'm also the founder of OuraNote, a mental health journaling app powered by an AI therapist, geolocation services, and an ABSA model that tracks a person's emotions throughout the year. On campus, I serve as the Technical Director and Secretary of the Data Science Society at UW where I mentor students, organize workshops and hackathons, conduct interviews, and guide members toward success. These experiences have shaped me into a well-rounded computer scientist and leader. I'm actively seeking Summer 2027 internships in Software Engineering, Web Development, AI/ML, and Research, where I can tackle real-world problems, learn from those around me, and deliver meaningful results.
• Analyzing information flow and attack surfaces in LLM-orchestrated agentic frameworks • Conducting literature reviews and analyzing research papers to identify weaknesses and attack surfaces in existing LLM-orchestrated agentic frameworks • Developing a structured framework to trace and reason about data propagation across agents, memory modules, orchestrators, and external tools • Built case studies analyzing how different system components enable attacks stemming from the proactive information-seeking behavior of AI agents • Currently co-authoring a Systematization of Knowledge (SoK) paper on attack surfaces in LLM-orchestrated agentic frameworks • Presented this research at the Paul G. Allen School Undergraduate & Master’s Research Showcase and was recognized as a runner-up among nearly 70 presenting research groups
• Developing and maintaining the website for UW’s Data Science Society. • Creating responsive and accessible UI using HTML, CSS, JS. • Showcasing projects, events, and blog content for students. • Managing GitHub repository and collaborating with team members. • Organized the DSML workshop and mentored 100+ students on resumes, interviews, and ML. • Led recruitment by interviewing and selecting candidates for technical and leadership roles within the society. • Organized speaker sessions, workshops, and technical events to support student learning and engagement. • Planned and organized Everybody Hacks, a full-day hackathon with 80+ participants across 3 tracks, coordinating sponsors, judges, certificates, logistics, and participant communication.
Co-founded and engineered OuraNote, a secure and expressive mental health journaling app designed for self-reflection, emotional wellness, and future planning. • Built and deployed a full-stack web application using React.js, Supabase, and Capacitor. • Implemented core features including daily diary entries, letters to your future self, encrypted vaults, and a personal timeline. • Designed an intuitive, calming user interface to foster private and meaningful journaling experiences. • Currently developing community-driven features to support safe and anonymous peer engagement. • Achieved early adoption with 30+ users onboarded during the initial launch. • Designed an AI therapist offering real-time emotional conversation and guidance. • Implemented AI-driven sentiment analysis to enhance journaling experience. • Designed and implemented an “Emergency Care Nearby” feature to help users quickly locate nearby therapy, crisis support, and mental health resources during moments of need.
The Startup Club is a student-driven community focused on innovation, collaboration, and entrepreneurship. We organize networking events, mentorship programs, and funding opportunities to help students turn their ideas into reality.
• Designed and implemented complex SQL queries to extract detailed asset, location, and financial data from Oracle Fixed Assets. • Built and deployed a REST API with 6 endpoints using Spring Boot (Java), enabling FlexiPro (internal asset management app) to replace manual query entry with automated data access. • Improved data accessibility and accuracy, significantly reducing manual effort and error-prone workflows for end users. • Tested and documented API functionality using Postman, ensuring smooth integration and reliability. • Developed a React-based frontend app that consumed the new API, providing users with a clean and intuitive UI for asset data visualization. • Enhanced system scalability and performance by creating a modular API service layer that streamlined frontend–backend communication.